|
J.UCS Special Issue on Dependability Evaluation and Validation
Klaus Echtle (Universität-GH Essen, Germany)
echtle@informatik.uni-essen.de
Jörg Keller (FernUniversität-GH Hagen, Germany)
joerg.keller@fernuni-hagen.de
Abstract: The field of dependable computing can be roughly
parted in the area of constructing hardware or software systems that
support dependable computing, and in the area of assessing such
systems. By assessment, we understand validation of proposed systems
and evaluation of implementations, both model based or experimental.
Research in this field covers many branches of computer science, such
as algorithms, software engineering, distributed computing,
communication systems, hardware design, and the broad field of
stochastic and deterministic modeling. The field as a whole is of
increasing importance as daily life more and more depends on computer
systems that work safely and reliable. This trend gives the
motivation for the special issue.
Within the field, the design of dependable systems has always drawn
more attention in publications, although the importance of assessment
- especially validation before costly implementation efforts start -
is growing. The challenge of assessment lies in the complexity of
fault-tolerant and safety-supporting systems, which requires special
efforts in various analysis methods: chosing the degree of details in
modeling, exploring large state spaces of models, provision of
realistic and yet tractable fault models, conducting fault injection
experiments in both the models and implemented systems (the latter for
model validation and parameter determination). These techniques affect
certification techniques as well, because certification nowadays is
done in parallel to the design and implementation process rather than
on the basis of the completed product only. Thus, we put our emphasis
on the aspects of evaluation and validation within the field of
dependability.
The six papers in this issue represent a broad spectrum of
activities within the field and cover both theory and practice.
Schneeweiss reports on modeling efforts that show alternatives to
classical methods like markov chains. Kochs et.al. apply different
modeling techniques in order to efficiently approximate reliability
parameters. Gärtner gives a survey of techniques to specify and
verify fault-tolerant (software) systems. Benso et.al. and Moreno
et.al. apply fault injection to hardware prototypes in order to verify
correct implementation of their fault-tolerance mechanisms. Finally,
Mauser and Thurner present a case study from an industrial setting. We
hope you enjoy the selection.
|